# Save 3D reconstruction for machine learning

**URL:** <https://discourse.slicer.org/t/save-3d-reconstruction-for-machine-learning/10096>\
**Category:** Support\
**Created:** [February 4, 2020, 2:00am UTC](https://discourse.slicer.org/t/save-3d-reconstruction-for-machine-learning/10096 "2020-02-04T02:00:26Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![manu1991](https://avatars.discourse-cdn.com/v4/letter/m/dfb087/32.png) [@manu1991](https://discourse.slicer.org/u/manu1991)\
**Post date:** [February 4, 2020, 2:00am UTC](https://discourse.slicer.org/t/save-3d-reconstruction-for-machine-learning/10096/1 "2020-02-04T02:00:26Z")

</div>

hi there,  
Im using Slice 3D for identification of the Urethra in prostate cancer. I’m making few sets of training data that I’ll export in python for the machine learning and deep learning. How should I save the format of the segmentation?

---

<div class="post-metadata">

**Author:** ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)\
**Post date:** [February 4, 2020, 4:25am UTC](https://discourse.slicer.org/t/save-3d-reconstruction-for-machine-learning/10096/2 "2020-02-04T04:25:12Z")

</div>

In general, I would recommend to use .nrrd format for storing image volumes (simple format that can properly save image geometry, it has human-readable header, and reader/writer in most development environment). You can use `slicer.util.saveNode()` to save a volume node to file.